Roy Sutton wrote:> Here''s the last piece of the puzzle you''ll need to get
this working
> (other than modifying the platform specific stuff like rakemswin.rb).
Looks cool, but I don''t think it''s quite ready for inclusion.
Hopefully
after we are using a "real" SCM tool it will be easy for you to
regenerate this changeset, and for me to re-apply it.
I think we either need to bite the bullet and include scintilla support,
or find ways of making it more optional and less intrusive. It looks
like the event callback stuff will be the most difficult in that respect.
Probably my biggest dislike was the pervasive testing of an environment
variable. Well, that''s how it felt, even if it was just one or two
places. I think the environment variable (if we go that route) should be
read by the rakefiles, and everything else should be handled by
command-line defines of WXSCINTILLA or somesuch.
I tend to hate environment variables, so if it''s going to remain
optional (and I''m not sure it should), we could either have the
existance of a file cause its inclusion, or we could have a config file
of sorts that developers could edit to include or exclude modules.
If nothing else, let''s read the environment variable once, and set a
global or something, so we can call a function like use_scintilla? to
decide whether or not to do the scintilla stuff in Ruby,
I''m still trying to figure out the big picture of all this. If we made
your patches non-optional, would wxScintilla headers be required on a
machine building wxruby2? After wxruby2 is built, could it run on any
machine, but would simply raise an exception if wxScintilla was called
but not present on the system?
Thanks for your work on this. One way or another, wxruby2 *will* support
wxScintilla.
Kevin